A Dynamic Mapping of the Creative Industries

5/21/2018 A Dynamic Mapping of the Creative Industries

1/79

A DYNAMICMAPPINGOF THE UKSCREATIVEINDUSTRIES

Hasan Bakhshi, Alan Freeman and Peter Higgs

This version January 2013


2/79

About Nesta

Nesta is the UKs innovation foundation. We help people and organisations bringgreat ideas to life. We do this by providing investments and grants and mobilising

research, networks and skills.We are an independent charity and our work is enabled by an endowment fromthe National Lottery.

Nesta Operating Company is a registered charity in England and Wales with company number 7706036 andcharity number 1144091. Registered as a charity in Scotland number SC042833. Registered office: 1 Plough Place,London, EC4A 1DE

www.nesta.org.uk

Nesta 2013.
http://localhost/var/www/apps/conversion/tmp/scratch_3/wwww.nesta.org.ukhttp://localhost/var/www/apps/conversion/tmp/scratch_3/wwww.nesta.org.uk


3/79

3 A DYNAMIC MAPPING OF THE UKS CREATIVE INDUSTRIES

EXECUTIVE SUMMARY

This paper argues that, despite its strengths, the UK Department of Culture, Media and

Sport (DCMS) classification of the creative industries contains inconsistencies which need

to be addressed to make it fully fit for purpose. It presents an improved methodology

which retains the strengths of the DCMSs approach while addressing its deficiencies. We

focus on creative intensity: the proportion of total employment within an industry that is

engaged in creative occupations.

Our analysis brings to light inconsistencies that undermine the strengths of the DCMS

definition as a de factoworld standard, and will detract from the understanding which it

has brought to the study of the creative economy, above all under conditions of structuraleconomic change, such as digitisation.

Using the list of occupations which DCMS treats as creative, the intensity of the industries

it defines as creative falls within a narrow range with only minor exceptions that

is on average over 25 times greater than in the rest of the economy. This is a defining

characteristic of such industries. However, DCMSs choice of industries excludes important

codes with high creative intensity that account for large amounts of employment.

In addition, DCMSs choice of occupations is itself open to question, because the criteria

by which they are classified as creative are not clear. We propose a rigorous method for

determining which occupations are creative, scoring all occupations against a grid of five

theoretically grounded criteria. The grid score of those occupations that DCMS considers

as creative also lies in a range significantly above the grid scores of other, non-creative

occupations. However, as with its choice of industries, DCMSs choice of occupations

excludes codes that account for significant employment and which, on the strength of a

rigorous classification, should be included. It also includes a small minority of codes which

should be excluded.

We then propose a fully consistent classification by using these occupations to identify,

on grounds of creative intensity, those industries that appear inappropriately included and

excluded in the DCMS industrial classification (our baseline). We conduct a sensitivity

analysis to show that this classification lays the basis for a robust and consistent selection

of industry codes. This accords with the reality, which should be squarely faced, thatuncertainty is a defining feature of emergent areas subject to persistent structural change

like the creative industries, and should be dealt with in a systematic way.

Our baseline classification suggests that the DCMS inappropriately excludes a large (and

growing) software-related segment of the creative industries. We argue that significant

numbers of new digital creative businesses in fact reside within this segment, reflecting

an increasingly tight interconnection between content production and its digital interface.

Our baseline estimates suggest that in its 2011 Statistical Release, the DCMS understated

the size of creative employment in the UK by 997,500 of which 460,000 falls within the

creative industries and 537,500 outside the creative industries.

Our estimates, like the DCMSs latest published estimates, are computed using the ONSs

SOC2000 classification of occupations. In 2013, the DCMS will adopt the Office for

National Statistics new SOC2010 classification which, in general, permits an improved


4/79


discrimination between which occupations are creative and which are not. We estimate

that the transition to SOC2010 will produce lower estimates of employment in the creative

economy by about 15 per cent.

Our baseline estimates show that creative economy employment is now a highly significant

and growing component of the workforce as a whole, accounting for 8.7 per cent of it

by 2010 as compared with 8.4 per cent in 2004. Our estimates also confirm a feature of

DCMSs estimates which has been documented in previous Nesta research: the majority

of creative workers are employed outside the creative industries in the wider creative

economy; this part of the creative workforce has grown particularly strongly, rising by 10.6

per cent between 2004 and 2010.

Our work shows that the creative industries do not rely, either wholly or mainly, on

traditional content or ICT activities alone. Rather, a new economic phenomenon has

emerged characterised by a parallel application, within single industries, of ICT and othercreative skills together. This strongly suggests that any attempt to separate ICT from

other creative work or to reduce the creative industries either to an offshoot of content

production, or for that matter a branch of the software industry, will not succeed. Thus our

sensitivity analysis includes, among other possible variants, the impact of removing the

main software occupation codes from the list considered to be creative occupations. Even

after this is done, ICT industries employing large numbers of people emerge as intensive

users of the remaining creative occupations. On this alternative scenario, the software-

related industries still contribute 213,000 jobs to the creative industries. The nonsoftware

creative industries are also very important employers of ICT labour.

We describe our approach as a dynamic mapping because a systematic method for

identifying the most creative industries produces a classification that does not over-

react to small fluctuations in the underlying data, but can respond to structural economic

changes. Intensity data can be used to compare like with like over time. We thus derive a

reasonably robust estimate of growth of creative economy employment which, between

2004 and 2010, rose by 6.8 per cent - more than five times the growth rate of the non-

creative workforce, measured on a comparable basis over the same period. In 2010, almost

2.5 million were employed in the UKs creative economy, of which 1.3 million worked in the

creative industries.


5/79


ACKNOWLEDGEMENTS

We would like to thank Mark Spilsbury for the invaluable advice he has given throughout

this research.

The statistics in this report are adapted from data from the Office for National Statistics

licensed under the Open Government Licence v.1.0.


6/79


1. INTRODUCTION

Considering it was introduced 14 years ago, the DCMSs (1998) classification of the

creative industries has arguably stood the test of time well. It has become a de facto world

standard. Creative industries estimates have become a regular feature of policy life and are

widely used and cited. The DCMS classification, in one form or another, has prevailed as the

preferred definition of the creative industries. This success strongly suggests that there is a

real economic entity which the classification captures, at least in part. It describes features

of the modern creative economy which are to be found in diverse countries throughout the

world, which are becoming more marked with the passage of time, and which correspond,

in some important respects at least, to the experience of the creative industries themselves.

This success masks a major shortcoming of DCMSs classification, however: it is

inconsistent. Although it does reflect an underlying economic reality, it does not fully

capture that reality; it excludes industries with the same features as the great majority of

those it includes, and includes others that do not share these general features, without a

clear rationale for doing so. In a fully consistent definition, by contrast, all industries in the

definition would share key common features, and no industry would be excluded, if it also

shared these common features.

A second problem is that the economic reality has itself changed, and the definition

has not been updated in line with these changes, notably digitisation and the fact that

increasing numbers of industries are embracing creativity as a way of gaining competitive

advantage. A key feature of DCMSs original definition (which informs its industry

classification) is its flexibility:

those industries which have their origin in individual creativity, skill and talent and which

have a potential for wealth and job creation through the generation and exploitation of

intellectual property.

This definition can accommodate change in principle. But this advantage has not been

exploited, and the actual industries and occupations considered to be creative are still

rooted in the conditions of the late 1990s.

The central obstacle to correcting these inconsistencies is that no explicit methodunderpins the DCMSs classification system. This lack of method only expresses a deeper

problem, which is that the concept of creativity itself was never defined. The often-

cited definition that we have just given is a policy guideline, not an analytic definition. It

offers a generalised rationale, but no explicit criteria for making informed judgements on

what should be counted as creative, and what should not. As such, it is not transparent;

decisions on which Standard Occupational Classification (SOC) or Standard Industrial

Classification (SIC) codes to include are not structured to permit informed discussion by a

community of practice including policymakers, practitioners and researchers. This contrasts

with the way that, for example, definitions of R&D and innovation have been developed

in such publications as the OECDs Frascatiand Oslomanuals, or cultural activity in

UNESCOs (2009) cultural statistics framework.1

This reflects a broader problem which is not of the DCMSs making: creativity is generally

speaking a poorly defined concept, and there is no agreed objective basis to judge what


7/79


is, or is not, creative. Writers casually impose their own definitions. Three defining works in

the field Florida (2002), Caves (2002) and Cox (2005) offer definitions which overlap

and, to a degree, mutually re-enforce each other, but certainly do not coincide. This is

not surprising, since each writer has one particular focus for example, Florida on theworkforce and its relation to urban space, Caves on the contractual structure of creative

business, and Cox on the relation of design to business innovation. However, though each

is interesting and valid in its own sphere, none addresses the wider question: what do we

mean by the word creativity? nor provides a definition of the creative industries rooted

in a systematic answer to that question.

Lacking a consistent, objective or transparent framework for selecting particular SIC

and SOC codes as creative and others not, we should not be surprised that the DCMS

has struggled to keep its classifications up to date in the face of structural changes such

as digitisation, and has retained internal inconsistencies, addressed in this paper, which

obstruct the production of reliable and trustworthy evidence.

The purpose of this paper is to address the shortcomings of the DCMS classification based

on a rigorous, analytic method which understands the creative industries as an integrated

economic whole. We are guided by three principles. First, the method should be robust;

the estimates to which it gives rise should not change by large amounts in response to

small changes in the underlying data or its classification. Second, it should be responsive:

capable of step by step adjustment to deal with structural, longerterm changes in the

economy. Third, it should be transparent: other analysts and researchers, with access to the

same data, should be able to reproduce its results. Such rigour is required not for arcane

reasons, but because a definition that matches economic reality will ensure that the wider

unity of practice, amongst those who use and produce creative industries statistics, is

regulated by a unity of understanding.


8/79


2. ORIGIN OF THE CREATIVE

INTENSITY METHOD

The method we use in this paper focuses on a measure which Freeman (2004: 7) termed

creative intensity, defined as the proportion of workers in any given creative industrythat

are engaged in a creative occupation. This approach draws on a key feature of the DCMS

classification: it includes a definition of both industries and occupations. This distinguishes

it from most other industrial classifications, including the SIC system itself, which define

only industries.2

The approach itself is rooted in the early work of the European Leadership Group onCulture (known as LEG), which informed the approach of the original DCMS (1998, 2001)

mapping documents. As Deroin (2011) explains, the development of European Working

Groups on cultural statistics began in November 1995, when the European Council of

Culture Ministers adopted the first resolution on the promotion of statistics concerning

culture and economic growth. This resolution invited the European Commission to

ensure that better use is made of existing statistical resources and that work on compiling

comparable cultural statistics within the European Union proceeds smoothly.

In response to this request, the Commission encouraged the creation of the first

European pilot working group on cultural statistics, known under the acronym LEG

Culture (Leadership Group on Culture). From 1997 to 2004, the LEG and its following

operational European working groups drew up the first European framework for cultural

statistics and developed specific methodologies such as the method for estimation of

cultural employment. (Deroin 2011:1).

This led in 2001 to a tool, developed by the European Task Force on cultural employment,

to produce a culture matrix which brings together cultural professions and cultural

activities. As Deroin (2011:15) explains:

This method for assessing cultural employment uses the results of the European Labour

Force Survey (LFS), which has the advantage of being based on a sample of households

in all the EU Member States (as well as in the candidate countries and the EFTA),

and of being structured around 2 reference classifications: the NACE which classifiesthe employers main activity, and the ISCO which classifies professions The method

consists in estimating all cultural employment in the economy, that is, employment in all

cultural activities along with cultural jobs in non-cultural activities. The estimate can be

performed by using two classifications (NACE and ISCO) used in the LFS. Once the most

refined posts are filled in, it is simple to make an estimate of cultural jobs:

Cultural employment =

cultural occupations (A)

+ non-cultural occupations in cultural activities (C)

+ cultural occupations in non-cultural activities (B)


9/79


The DCMS (1998) classification reproduced the core idea that creative and cultural activity

is best captured by describing, and measuring, both the industries whose outputs may

be considered creative, and the occupations whose activities may be considered creative.

But DCMS did not draw any special connection between the two. Rather, it regarded thecreative occupations as an additional component of creative employment as a whole,

simply adding creatively-occupied workers outside the creative industries to those inside,

and even assigning them to industries they did not work in. It paid little attention to

the specialised use which the creative industries made of their creative talent; it did not

until 2011 publish statistics recording the number of creative workers that work withinthe

creative industries and has not really paid any systematic attention to this aspect of its own

statistics.

Three groups of researchers have drawn attention to the distinctive role of the creative

workforce inside the creative industries themselves. Peter Higgs and Stuart Cunningham,

working at the Centre of Excellence for Creative Industries (CCI) at Queensland Universityof Technology, devised an approach they termed the Trident method (Higgs et al., 2005).

Using a terminology we employ throughout, they called creative occupations inside the

creative industries specialist jobs and those outside the creative industries embedded

jobs: they coined the term support jobs, now adopted by DCMS, to describe the additional

jobs within the creative industries which were not themselves creative occupations.

Working independently, Freeman (2004:7) began producing measures of creative

intensity and showed that this was systematically higher in the creative industries than

elsewhere, was increasing over time, and was particularly high in London and the South

East of England. Nesta encouraged the development of these ideas in the UK, leading to a

number of publications on the creative industries that focussed on the role of the embedded

workforce (Higgs et al., 2008), Bakhshi et al., 2008).3Freeman (2008b:15) concluded that:

If we think of this labour as a resource, and the sectors outputs as a product, then it

begins to make sense to conceive of the industry as a specialised branch of the division

of labour which uses this resource to produce specialist products.

We can illustrate this by asking the simple question: where are creatively occupied workers

actually employed? Table 2.1 provides a basic breakdown for the industries and occupations

defined by DCMS as creative. In this Table, the components of creative employment are

highlighted. The 476,800 jobs in the first row and column are the specialist jobs and

the 600,900 in the first column and second row are the embedded jobs. The remaining

420,500 in the second column and first row are the support workers.

The results are qualitatively very significant. 53 per cent of those employed within the

industries which DCMS defines as creative are engaged in occupations which DCMS defines

as creative. This is over 25 times higher than in those industries that DCMS does not define

as creative. It is also consistent across nearly all the DCMS industries. As Table 2.2 shows,

only three of the eleven DCMS sectors defined by industrial codes have intensity lower

than 35 per cent. Moreover the low intensity recorded for the sectors 8 and 12 (Software/

Electronic Publishing and Digital and Entertainment Media) is entirely a consequence of the

reclassifications introduced with DCMSs 2011 Statistical Release. If these reclassifications

had not been made, the intensity in these two sectors combined would have been 58 per

cent.


10/79


TABLE 2.1:EMPLOYMENT IN THE CREATIVE INDUSTRIES

Source: Creative Industries Economic Estimates Full Statistical Release, 8 December 2011, page 28

TABLE 2.2:INTENSITIES IN THE DCMS SECTORS, 2011 ESTIMATES

Table 2.3 provides a more detailed view, looking at the individual SIC codes which are

used in DCMSs classification in the above sectors. Intensity within five-digit SIC codes

cannot be determined with any more accuracy than for four-digit codes, since the

Labour Force Survey (LFS) section of the Annual Population Survey (APS) the basis of

DCMSs estimates only provides data classified at the four-digit SIC level. From now on

we therefore refer to the four-digit codes, within which the firms classified by DCMS as

creative are to be found, unless the contrary is stated.

Occupation

Industry Creative Other Total in this Intensity (CreativelyOccupations Occupations industry Occupied/Total

Employment in theIndustry)

Creative Industries 476,800 420,500 897,300 53%

Other Industries 600,900 27,622,800 28,223,700 2%

Total in this occupation 1,077,700 28,043,300 29,121,000 4%

Creative Other Total IntensityOccupations Occupations

1. Advertising 45,900 69,400 115,300 40%

2. Architecture 67,300 36,200 103,500 65%

3. Art & Antiques 500 8,300 8,800 6%

5. Design 56,400 42,100 98,500 57%

6. Designer Fashion 3,700 2,900 6,600 56%

7. Film, Video & 28,700 29,500 58,200 49%Photography

9&10. Music & Visual and 138,400 52,800 191,300 72%Performing Arts

11. Publishing 71,300 111,500 182,700 39%

8&12. Software/Electronic 900 22,300 23,200 4%Publishing

8&12. Digital & 2,000 11,200 13,200 15%Entertainment Media

13. TV & Radio 61,700 34,200 96,000 64%

Total 476,800 420,500 897,300 53%


11/79


Only one of the four-digit industries identified by DCMS as containing creative industries

has an intensity lower than the national average for the economy as a whole. More than half

these codes accounting for 75 per cent of those working within these industries have

intensities greater than 30 per cent.

TABLE 2.3:CREATIVE INTENSITIES IN CODES DEFINED BY DCMS AS CREATIVE

Industry Intensity

9003 Artistic creation 90%

5912 Motion picture, video and television programme 89%post-production activities

9001 Performing arts 81%

6010 Radio broadcasting 73%

7420 Photographic activities 73%

5911 Motion picture, video and television programme 68%production activities

7111 Architectural activities 65%

6020 Television programming and broadcasting activities 56%

5814 Publishing of journals and periodicals 55%

7410 Specialised design activities 55%

9002 Support activities to performing arts 52%

7312 Media representation 47%

5920 Sound recording and music publishing activities 40%

5813 Publishing of newspapers 39%

7311 Advertising agencies 39%

1820 Reproduction of recorded media 36%

5811 Book publishing 34%

5819 Other publishing activities 28%

9004 Operation of arts facilities 23%

1813 Prepress and media 20%

5913 Motion picture, video and television programmedistribution activities 19%

5914 Motion picture projection activities 13%

6201 Computer programming activities 11%

1411-1520 Clothing and accessories 7%

4779 Retail sale of secondhand goods in stores 6%

4778 Other retail sale of new goods in specialised stores 5%

5829 Other software publishing 3%


12/79


This is not a general feature of the relation between industries and occupations.

Table 2.4 shows the intensities of the main occupational groups within the main industrial

Sections, as defined by the Office for National Statistics (ONS). These Sections definethe standard groupings which the ONS, applying international standards, considers to

be industries properly defined. Very few of the occupational intensities match even the

average seen in the creative industries. The highest is Teaching and Research Professionals,

perhaps the most specialised occupation that is explicitly defined, and accounts for 45 per

cent of the workforce in the Education sector significantly lower than many intensities

found in the creative industries. Science and Technology Professionals, an occupation

which might be expected to show high degrees of industrial specialisation, given their

similarity to the creative industries in other respects, do not form a particularly high

proportion of the workforce of any industrial Section.

Moreover in many cases where intensities even approach those found in the creativeindustries, we find that the occupations concerned are intensively employed across a range

of industries, unlike the creative occupations which tend to be heavily concentrated in

the creative industries and dispersed in the others a point we will shortly investigate in

more depth. So, for example, Corporate Managers make up 28 per cent of employment

in Financial and Insurance Activities. But they are clearly a general resource used in a wide

range of industries, showing intensities of 19 per cent in Manufacturing and 17 per cent in

Electricity, Gas etc. We should expect that any large occupational group will be intense

across a range of industries. The peculiarity of the creative industries is that the high

intensities apply only to a quite specific group of industries, that employ these types of

workers in much higher proportions than do almost all other industries.

It is possible that these low intensities reflect an inappropriate selection of occupations

so that, if occupations were judiciously chosen, as with the creative occupations, they

would account for a higher proportion of employment in certain industries that are in some

sense their natural home. Thus Skilled Construction and Building Workers account for

39 per cent of employment in Construction (a not unsurprising statistic), which is itself

low when compared with the intensities typical of the creative industries, but we might

suppose that this proportion would increase if we added other occupations that were

also intense within construction. But there are no obvious such groups the next highest

intensities are Corporate Managers, and Skilled Metal and Electrical Trades both of which

are of the dispersed, general type discussed above.
http://www.bbc.co.uk/news/10604117


13/79


TABLE 2.4:OCCUPATIONAL INTENSITIES IN THE STANDARD GROUPS

Corporate Managers

Managers andProprietors inAgriculture andServices

Science andTechnology

Professionals

Health Professionals

Teaching andResearchProfessionals

Business and PublicService Professionals

Science andTechnology AssociateProfessionals

Health and Social

Welfare AssociateProfessionals

Protective ServiceOccupations

Culture, Media andSports Occupations

Business and PublicService AssociateProfessionals

AdministrativeOccupations

Secretarial andRelated Occupations

Skilled AgriculturalTrades

Agriculture

,ForestryandFishing

Miningand

Quarrying

Manufactur

ing

Electricity,

Gas,SteamandAirConditioning

Supply

WaterSupply

Constructio

n

Wholesale

andRetailTrade

TransportationandStorage

Accommod

ationandFoodServices

Information

andCommunication

FinancialandInsuranceActivities

RealEstate

Activities

Professiona

l,Scientificand

TechnicalActivities

AdministrativeandSupportServiceActivities

PublicAdm

inistrationandDefence

Education

HumanHea

lthandSocialWork

Arts,entertainmentandRecreation

OtherServiceActivities

17

1

14

0

0

3

5

0

0

0

6

9

1

0

13

4

6

0

1

2

4

0

0

0

5

8

1

0

15

1

5

0

0

2

2

0

0

0

2

5

2

0

3

10

1

0

0

0

0

0

0

0

1

3

2

45

15

2

11

0

0

3

6

0

0

0

5

6

1

0

15 10 3 25 28 12 18 12 13 3 8 6 7

4 1 16 3 1 19 2 5 1 1 1 9 6

1 1 0 26 5 0 10 1 3 1 1 1 2

1 0 0 0 0 0 1 0 1 0 8 0 0

0 0 0 0 0 0 1 1 2 45 1 1 1

0 1 0 3 6 4 20 1 7 1 4 3 10

0 1 0 4 2 1 4 1 2 2 1 1 2

1 0 0 0 0 7 0 0 4 1 24 0 3

0 0 0 0 0 0 0 0 17 0 0 0 0

0 0 0 15 0 0 8 1 1 2 0 28 2

4 5 1 5 24 15 11 7 10 3 3 4 3

6 9 2 5 23 15 10 9 25 4 6 10 7

1 1 2 1 2 5 5 2 3 3 5 3 3

0 0 0 0 0 0 0 8 0 0 0 5 0

19

1

9

0

0

1

4

0

0

1

4

6

1

0


14/79


Skilled Metal andElectrical Trades

Skilled Constructionand Building Trades

Textiles, Printing andOther skilled Trades

Caring PersonalService Occupations

Leisure and OtherPersonal ServiceOccupations

Sales Occupations

Customer ServiceOccupations

Process, Plant andMachine Operatives

Transport and MobileMachine Drivers andOperatives

Elementary Trades,Plant & StorageRelated Occupations

ElementaryAdministration andService Occupations

12

7

1

0

0

4

10

5

2

1

1

4

1

0

0

0

1

2

9

16

5

17

10

39

1

0

0

0

0

5

3

7

1

2

0

1

1

0

0

0

2

3

25

1

13

1

0

0

0

1

0

21

11

1

2

14

3

5

0

0

1

1

18

3

8

1

5 3 0 4 0 1 1 2 1 0 0 1 5

0 0 0 0 0 2 1 0 0 0 0 0 1

2 0 13 1 0 0 0 0 0 1 1 1 2

0 0 0 0 0 2 1 1 2 20 33 2 4

0 4 2 0 0 2 0 5 0 2 1 7 31

36 1 4 2 3 7 1 2 0 0 0 2 1

2 2 1 2 5 2 1 5 1 0 0 2 1

2 1 1 0 0 0 1 1 0 0 0 1 1

4 40 1 0 0 0 0 3 1 1 0 0 1

5 8 1 1 0 1 0 4 1 0 0 2 1

7 12 48 2 1 3 1 30 3 8 4 10 5

Note:Highlighted cells show intensity greater than 10 per cent

The special role of the creative workforce within the creative industries has led all three

groups of researchers that we previously identified to agree that a defining feature of

the creative industries is its workforce, and in particular the special use that they make of

particular types of workers. The working assumption that informs this paper is that the

creative industries are a specialist branch of the division of labour that has discovered

how to harness the capabilities of this workforce to produce outputs which, it turns out,

constitute a growing share of the value added in most advanced economies.

The justification for this assumption goes beyond the evidence of the intensity figuresalone. A range of research suggests that the pragmatic validity of the creative industries

arises because they perform a definite and growing economic functionwhich arises from

fundamental changes in society, the most central being digitisation, the rise of the content

industries, and the steadily growing share of discretionary spending in total economic

demand. This idea is developed at greater length, and justified, below in Section 4, where

we deal with the definition of creative occupations.


15/79


3. PROBLEMS AND

INCONSISTENCIES INTHE DCMS CLASSIFICATIONS

Chart 3.1 summarises the way this guides our approach, presenting creativelyoccupied

jobs as a frequency distribution. The vertical axis shows the creatively-occupied jobs within

industries having the intensities shown on the horizontal axis.

CHART 3.1:DISTRIBUTION OF CREATIVELY-OCCUPIED JOBS BY CREATIVEINTENSITY

The chart shows how creatively-occupied jobs are distributed between industries. Thehorizontal axis shows ten bands of increasing intensity, the smallest covering zero to 5per cent and the largest covering 85 to 95 per cent. Each column shows the creativeemployment accounted for by the industries whose intensity falls within that band: thusthe 22,800 creatively-occupied jobs within code 6201 (Computer Consultancy) in whichintensity is 11 per cent, will be counted within the bar over the band 05-15 per cent. Thenumbers inside the bars show the number of industries that fall within this frequency range.

This is a bimodal distribution with two peaks around which intensity is clustered one

which appears to lie between 0 and 15 per cent, and the other between 65 and 75 per cent.

150,000

100,000

50,000

200,000

250,000

300,000

350,000

0

00-05 05-15 15-25 25-35 35-45 45-55

Creative Intensity, Per Cent

Number ofcreatively-

occupiedjobs

Creative intensity, per cent

55-65 65-75 75-85 85-95

458 84 31 14 6 4 6 6 2 2


16/79


We can study this in more detail by asking how much employment, in each frequency band,

is accounted for by SIC codes that are included, at least in part, in the DCMS classification,

and SIC codes that are not. This is shown in Chart 3.2.

This clearly confirms that a group of industries are distinguished by a markedly higher

tendency to employ creative workers. But it also points to misallocations in the DCMS

statistics: a definite group of industries which DCMS does not treat as creative exhibit high

intensities, showing as a blip in the distribution of the non-creative industries peaking

at 55-65 per cent. In addition, a significant number of industries that DCMS classifies as

creative exhibit intensities well below the average for the creative industries.

CHART 3.2:DISTRIBUTION OF CREATIVELY-OCCUPIED JOBS BY CREATIVEINTENSITY, PARTITIONED INTO DCMS-CREATIVE AND NON-DCMS-CREATIVE

The inconsistency becomes clearer, as do the possible means to correct it, if we restore

to the occupations considered creative the two software occupation codes which DCMS

dropped in 2011 but included in its 2010 estimates, these being IT Strategy and Planning

Professionals (2131) and Information and Communication Technology managers (1136).

We can then recalculate the intensities that result, giving Chart 3.3. On the one hand,

significantly fewer DCMS-creative industries exhibit the low intensities shown in Chart

3.2; particularly those lower than 25 per cent. But in addition, a much larger group of

nonDCMScreative industries now exhibit intensities above the average for the creative

industries, showing up as a new and larger blip between 45 per cent and 55 per cent,

dwarfing the blip between 55 per cent and 65 per cent which still remains. The first blip

includes the software-related industrial code 6202 with a creative intensity of 47 per cent,

150,000

100,000

50,000

200,000

250,000

300,000

350,000

0

00-05 05-15 15-25 25-35 35-45 45-55



occupiedjobs


55-65 65-75 75-85 85-95

DCMS-Creative Non-DCMS-Creative


17/79


employing 201,800 workers of whom 94,000 are creatively-occupied. Industrial code

6209 with an intensity of 29 per cent contributes a further 35,000 jobs of which 10,000

are creatively occupied. Finally, although the code 6201 with intensity of 34 per cent is

presented in Chart 3.3 as creative in its entirety (employing 207,000 of whom 70,000are creative), DCMS in fact only counts a small proportion of the employment from this

software-related code.

This suggests that a combination of creative skills across a spectrum of activities

contributes to the creative industries as a coherent grouping of sub-sectors. The growing

use of ICT in virtually all spheres of creative work suggests that creative talent has great

economic impact when working in tandem with ICT.

CHART 3.3:DISTRIBUTION OF CREATIVELY-OCCUPIED JOBS BY CREATIVE

INTENSITY, PARTITIONED INTO DCMSCREATIVE AND NON-DCMSCREATIVE, WHEN TWO EXCLUDED SOFTWARE OCCUPATIONS ARERESTORED

150,000

100,000

50,000

200,000

250,000

400,000

300,000

450,000

350,000

0

00-05 05-15 15-25 25-35 35-45 45-55

Number of

creatively-occupied

jobs

55-65 65-75 75-85 85-95

Creative Intensity, Per CentCreative intensity, per cent


This is confirmed if we add to the list of creative occupations a further software related

occupation code which DCMS has never treated as creative, namely Software Professionals

(2132). This gives rise to Chart 3.4 in which the distinctiveness of the two distributions

involved is particularly clear.

Employment in the non-DCMS-creative industries lies on a distribution skewed towards

zero, with two-thirds of all creatively-occupied jobs located in industries whose intensity is

less than 15 per cent. Employment in the DCMS-creative industries lies on a very different


18/79


distribution with 60 per cent of all creatively-occupied jobs located in industries whose

intensity is greater than 55 per cent.

This distribution provides further confirmation that that the DCMS selection of industriesinvolves misallocations; a large amount of creative employment, in industries that DCMS

does not treat as creative, resides in industries with an intensity (defined, as stated above,

on the basis of intensities that include all ICT occupations) in excess of 65 per cent.

A fact deserving especial attention is that the inclusion of ICT occupations significantly

modifies the distribution of intensity, and that their complete exclusion leads to the

much less coherent distribution of intensities seen in the DCMS classification in Chart

3.2. This points to a distinctive feature of the creative industries, which is their tendency

to use labour from software occupations and more broadly from ICT occupations in

combination with other forms of creative labour. This requires attention precisely because

of the structural changes to the creative industries brought about by digitisation, and moregenerally the impact of ICT, a point made by Nicholas Garnham (2005).

CHART 3.4: DISTRIBUTION OF CREATIVELY-OCCUPIED JOBS BY CREATIVEINTENSITY, PARTITIONED INTO DCMSCREATIVE AND NON-DCMSCREATIVE, WHEN SOFTWARE PROFESSIONALS (2132) AREINCLUDED

100,000

200,000

300,000

600,000

400,000

700,000

500,000

0

00-05 05-15 15-25 25-35 35-45 45-55


occupiedjobs

55-65 65-75 75-85 85-95



A comprehensive study of the role played by ICT, and software in particular, in the

transformation of the creative industries deserves to be the subject of further research.

It is complicated by the fact that the ICTbased industries are highly developed in other


19/79


fields too for example, in commerce and financial service industries, in the automation

of manufacture, in sciencebased industries, engineering and so on. Thus, the mere

employment of ICT talent is not always in itself an indicator of creativity. However, ICT

labour appears to play a special role within the creative industries, when it is deployed incombination with other types of creative labour. Table 3.1 therefore shows the intensity

of employment, within those industries that are already identified as intensive users of

other types of creative labour, of the three ICTrelated occupations we have discussed

above. These are, to recall, IT Strategy and Planning Professionals (2131), Information and

Communication Technology Managers (1136), and Software Professionals (2132).

The Table looks at intensity using nonICT creative occupations only (that is, those

occupations used by DCMS in its 2011 update), dividing all industries as before into two

groups: those that DCMS defines as creative and those that it does not. It then asks how

much of the additional employment that these industries provide consists of workers in ICT

occupations.

As Table 3.1 shows, within those industries that employ nonICT creative labour more

intensively than 10 per cent, 86 per cent of all ICT labour is employed in the DCMScreative

industries.

This confirms the economic rationality ofthe original DCMS classification, in both of which

software occupations figure among the mix that is treated as creative, leading to a more

consistent relation between industries and occupations than in the 2011 statistical release

and confirming the hypothesis that an essential characteristic of the creative industries is

the way that ICT creative occupations work with nonICT creative occupations within them.

TABLE 3.1 EMPLOYMENT OF ICT OCCUPATIONS IN INDUSTRIES THAT USE NONSOFTWARE LABOUR INTENSIVELY

*Exact figure suppressed due to disclosure control restrictions

Range ofintensity for Of which in Of which in Working in Working innonsoftware Total ICT creative noncreative creative non creativeoccupations employment industries industries Industries Industries

0005 539,900 12,900 527,000 2% 9.3% 2.4%0510 80,000 900 79,100 1% 0.5% 3.3%

1020 138,500 118,600 19,900 86% 49.6% 1.4%

2030 4,600 4,000 600 87% 5.4% 0.2%

3040 9,700 7,900 1,800 82% 4.3% 1.5%

4050 3,000 2,900 95%* 8.4% 20 27,500 24,900 2,600 91% 3.0% 0.6%

Additional ICT employmentwithin this intensity range

Proportion of ICT labour

ICT intensity innoncreativeindustries
http://localhost/var/www/apps/conversion/tmp/scratch_3/rationality.ofhttp://localhost/var/www/apps/conversion/tmp/scratch_3/rationality.of


20/79


This analysis thus confirms empirically that the creative industries as originally conceivedof by DCMS are economically distinct, and are distinguished by a markedly highertendency to employ creative workers, and that within this there is a strong tendency toemploy workers in ICT occupations in tandem with other creative occupations. This leads

us to conclude that intensity, including intensity of use of at least some ICT occupations,is a significant discriminator of industry creativity. If we are looking at an industry andattempting to judge whether or not it may be creative, the first port of call is to ask howfar it lies within the upper distribution shown in Charts such as 3.2, 3.3 and 3.4. We nowproceed to develop the above empirical insights into a rigorous definition.


21/79


4. A FIRST STEP TO A

SOLUTION: DEFININGCREATIVE OCCUPATIONS

The first problem is that the creative occupations which underlie the DCMS classification

are not themselves defined rigorously. In constructing Table 2.3, we employ the official

DCMS definition of creative occupations. But the industrial codes could be appearing in

that Table simply because the underlying occupations are wrongly defined as creative, and

others could be absent for the converse reason. Hence the need to look at occupations

more closely.

In this Section we attempt to define more rigorously what makes creative occupations

creative. In addressing this question, we return to the idea that the creative worker is

a decisive resource for the creative business. What is the economic role of the creative

worker? We can think of any productive activity as a sequence which passes from inputs,

transforms them in some process more or less specific to the industry, and produces

outputs as a result. This suggests that the way to conceptualise what a creative worker

does is to ask what does she or he contribute to theprocessthat produces the outputs

from the inputs?

To contextualise this, we return to the economic model of the creative industries which

informs this paper and was briefly introduced earlier.

Digitisation, and more generally ICT, provides the capacity to transcend the traditional

barriers of service production. These technologies facilitate the reproduction of a growing

range of services at any distanceby means of transmission technology, at anytime by

means of recording technology, and in anyquantityby means of copying and reproduction

technology. These lay the technological basis to deliver products and services which were

at one point confined to direct person-to-person contacts, to a far wider audience than

previously.

This has been accompanied by a parallel growth in creative content and service industries

that produce what is delivered through the new technologies. The relationship appearscomplex if the economic mechanism is not understood.4Paradoxically, for example, it has

also led to increased popularity of live performance, attendance at exhibitions, and so

on. Page (2007), for example, has consistently tracked, using royalty data, how consumer

spending on live music performance has increased.

At the same time, there has been a continued rise in spending on such products as fashion,

in which questions of taste and subjective perception of experience predominate over

pure quantity.5It is logical to view this as an outcome of the broader rise of discretionary

spending. In 1994, for the first time, UK families spent more on leisure products and

services than on food. By 2004 they were spending twice as much. Similarly, businesses

are investing more on creative services, such as design, advertising and software, than

other more tangible expenditures.6


22/79


These trends can be understood as a substitution effect: as digitisation has cheapened

creative products, consumers and businesses have increased how much they spend

on them. But, arguably, they also suggest that consumers and businesses increasingly

discriminatein their purchases, placing the highest premium on the most authentic anddirect experiences.7

The creative industries have responded to all of these new opportunities by developing,

to a high degree, the capacity to differentiate their products; to cater precisely for the

discretionary requirements of more segmented groups of clients or customers. This also

brings into play non-IP methods for realising the value-added supplied by the creative

process such as first-mover advantage most obvious in the fashion industry but

increasingly common elsewhere in which the seller, rather than placing a high emphasis

on management of copyright or patents, creates and maintains a client base on the basis

of brand, distinctiveness and novelty.

This requires a new form of production in which the key requirement is no longer the

production of large volumes at low prices, but a continuous succession of small runs of

products each varying from its predecessors and the competition in respects which

may be small, but are sufficiently adapted to customer needs, and sufficiently highly-prized

aesthetically otherwise, to attract the loyalty of a discriminating clientele.

In order to achieve this, the creative industries have become primary users of a specialist

workforce that knows how to satisfy the needs of a discriminating customer base. Our

interpretation of the different characteristics of this workforce are discussed later when we

undertake a more rigorous definition of it, but together they focus on the capacity to meet

what we term in line with common parlance in linguistics and computing requirements

expressedsemanticallyrather than in terms ofprocess. That is, the creative worker has a

concept of what kind of effect is desired, but is not told how to produce that effect in the

same way that, say, an assembly line worker or even skilled technician is instructed. The

creativity, in our view, consists in devising an original way of meeting a differentiated need

or requirement that is not expressed in precise terms.

This confers a unique and important quality on the creative worker within the creative

process, namely that it is difficult to mechanisethe creative process and hence to

substitute machines or devices for the humans, reversing a trend that has dominated much

of history. Implementing a creative decision is not really a creative role, we would argue,

but making one is. Technology has largely done away with the need for the highlyskilled

roles of typesetters and photo touchup artists. The former is now subsumed into the pagemanagement applications and style guides applied by art directors and graphic designers.

The traditional photo touchup artists palette of complex specialist physical techniques

such as dodge and burn are now plugins to applications such as Adobe Photoshop used,

again, by graphic artists.

In hindsight, while these crafts were important to the creative output of advertising,

they arguably were not themselves creative occupations. The continual process of

democratisation of technology lowers the cost and the technical skill needed to do

previously highly complex, but essentially non-creative, tasks. Editing a film is a creative

task but operating a 6 plate 35mm Steenbeck editing table under the direction of

the editor is not. The onset of digitisation has allowed the film director to make, and

implement, creative decisions directly, using programmes such as Avid or Final Cut Pro on

her or his laptop, or in a nonlinear editing suite, steadily eliminating dependence on purely


23/79


repetitive craft skills. Creatives adopt, adapt and absorb new technologies in pursuit of

creative excellence. They are seldom made redundant by it.

These workers are also engaged in specific and new types ofprocess, described by suchwriters as Caves (2002) and Chesborough (2003) which suggest additional indicators of

creativity in the economy. These include pre-market or gatekeeper selection mechanisms

(for example galleries, agents, distributors or publishers), project-based or open

collaborations (Caves describes this as the motley crew principle), contracts that manage

uncertainty rather than risk (which Caves terms the nobody knows principle) and so on.

A final important characteristic is the strong tendency towards geographical clustering at

a microspatial level, leading to such phenomena as West Londons film, broadcasting and

advertising clusters, or the Shoreditch Triangle.

These considerations inform an economic model of the creative industries; they may

be thought of as an industry, in the normal economic sense of the word, which has acharacteristic input, a characteristic output and a characteristic process of production,

through which the inputs are deployed to produce the outputs. The defining feature of the

creative industries does not lie, according to this approach, solely in producing cultural

outputs or in innovation or originality these are the province of other industries also.

It lies in their use of the workforce within a specific process to produce the outputs in

which these industries specialise. Their most unusual feature is that their distinctive input

is a type of labour creative talent. They are thus different, for example, from traditional

manufacturing industries which are defined either by physical, non-human or mechanical

inputs or outputs, or by mechanical processes: agriculture creates products from the land,

whilst manufacturing creates products that require machinery, and so on. They are defined

in summary by:

1. A common type of input or resource (the creative workforce).

2. Common features of the output (emphasis on content, product differentiation,shorter, often smaller, production runs, preponderance of cultural or culture-relatedoutputs, sale to discretionary markets, exploitation of both traditional IP and first-mover advantage).

3. Common processes of production (pre-market selection, uncertainty-managementcontracts, just-in-time short-run production methods, open innovation with anemphasis on collaborative contracts, geographical clustering at the micro level, andso on).

The workforce constitutes the link between all the above three aspects. Creative talent is tothe creative industries what the land is to agriculture or the machine to manufacturing: itsdefining indicator. It is a specialist resource that is used precisely because it knows how toimplement the processes and produce the results.

All these features have been recognised to a greater or lesser degree in previous research

which tends however to concentrate on one aspect of this economic model at the expense

of the overall picture. Thus as we have noted, Richard Florida focuses on the workforce and

clustering, Richard Caves on the nature of the creative contract, whilst yet others focus

centrally on the output of the creative industries. Yet none (Caves comes closest) really

consider the industries as a whole, taking into account the resources and inputs that they

deploy, the process in which they use them, and the outputs that result, and understanding

the relationship between these dimensions of production.


24/79


This is particularly evident in the otherwise excellent work that UNESCO (2009) has done

on defining the nature of cultural activities, which is centrally concerned either with the

end result of cultural activity for example, cultural consumption and participation or

in the interface between one producer and another, as in the analysis of the cultural valuechain. Yet other approaches, such as the definition offered by WIPO (2004, 2012) which

confines itself to intellectual property outputs, exclude other forms of capitalising on

differentiated output such as first-mover advantage, discussed above. This leads to one-

sided appreciations of what the creative industries actually do, to which, in our view, the

key is our understanding of the resource which makes them what they are: their creative

workforce.

This now allows us to give rigorous meaning to the idea of creative occupation. We define

this as:

These creative skills involve a combination of original thought all creative skills involving

problem solving to a greater or lesser degree with processes defined by collaborative

relationships to deliver or realise the output. We operationalise this definition by breaking it

down into a set of five criteria:

1. Novel process Does the role most commonly solve a problem or achieve a goal,even one that has been established by others, in novel ways? Even if a well-definedprocess exists which can realise a solution, is creativity exhibited at many stages ofthat process?

2. Mechanisation resistant The very fact that the defining feature of the creativeindustries is their use of a specialised labour force shows that the creative labourforce clearly contributes something for which there is no mechanical substitute.

3. Non-repetitiveness or non-uniform function Does the transformation which theoccupation effects likely vary each time it is created because of the interplay of

factors, skills, creative impulse and learning?

4. Creative contribution to the value chain Is the outcome of the occupation novel orcreative irrespective of the context in which it is produced; one such context beingthe industry (and its standard classification) of the organisational unit that hostsor employs the role? For example, a musician working on a cruise ship (a transportindustry) is still creative while a printer working within a bank is probably operatingprinting technology and hence would be considered mechanistic and not creative.

5. Interpretation, not mere transformation does the role do more than merely shiftthe service or artefacts form or place or time? For instance, a draughtsperson/CADtechnician takes an architects series of 2D drawings and renders them into a 3D

model of the building. While great skill and a degree of creative judgement areinvolved, arguably the bulk of the novel output is generated by the architect and notby the draughtsperson.

a role within the creative process that brings cognitive skills to bear to bring about

differentiation to yield either novel, or significantly enhanced products whose final

form is not fully specified in advance


25/79


Of course, each of these five criteria are problematic when considered in isolation, and

they do not offer hard and fast rules for determining whether an occupation is or is

not creative. There are also connections between them: it is unlikely that the activities

of someone who is constantly called on to devise new processes, to carry out newtransformations and to construct creative interpretations of their raw material can easily

be mechanised. But occupations which score positively on all or most of the indicators,

we believe, are very likely to function as an economic resource that the creative industries

require.


26/79


5. STEP TWO: RATING THE

STANDARD OCCUPATIONSUSING THE CREATIVE GRID

We applied the criteria established above to the Standard Occupational Classifications. All

occupations were examined and the value 1 assigned where the occupation complies with

the criterion, and 0 where it does not. The values were then totalled to provide an overall

grid score. We set a threshold of four to qualify an occupation as creative.

In this paper, we apply the ratings to the SOC2000 occupations so that we can comparethe results with the creative occupations in the latest published DCMS Creative Industries

Economic Estimates which also use SOC2000, and because data based on the SOC2010

classification is only available for the LFS from 2011 onwards (both our results and those

that DCMS has published so far, only go up to 2010). Unlike the SIC codes in the DCMS

definition, the SOC codes have changed relatively infrequently (only once during the life

of the estimates, in 2000 as the name suggests). This means that, although the industrial

classifications have breaks and discontinuities which make it difficult to deduce long-

term time trends, estimates of the total creatively occupied workforce provide a more or

less continuous time series since the year 2000. Whilst this does not solve the principal

problem of determining in which industries this workforce is actually deployed to produce

creative products, it is a useful anchor; for this reason we suggest that changes in the

occupation codes included in the classification should be changed relatively infrequently

and the transition to SOC2010 should be undertaken with an eye to continuity. In this

analysis, SOC2000 applies throughout, except in our final Section 12 in which we assess the

likely impact of the transition to SOC2010 on our estimates.

Table 5.1 shows the occupations which on this basis we select as creative, defined as

scoring 4 or 5 out of the possible total of 5. Table 5.2, for comparison, lists codes which

DCMS treats as creative but which we score less than 4 and which are therefore not

included in our final list.

Applying the Creative Grid produces a significantly higher total of creatively-occupied jobs

than the DCMSs selection. The differences are summarised in Table 5.3, which lists codesidentified as creative according to the Creative Grid but not recognised as creative by the

DCMS, and Table 5.4, which lists codes that DCMS counts as creative but which are not

grid-scored as creative.


27/79


SOC

code Occupation

1132 Marketing and sales directors 5 1 1 1 1 1

1134 Advertising and Public Relations managers 5 1 1 1 1 1 Yes

2131 IT Strategy and Planning professionals 5 1 1 1 1 1 Yes

2132 Software professionals 5 1 1 1 1 1

2431 Architects 5 1 1 1 1 1 Yes

2432 Town Planners 5 1 1 1 1 1 Yes

2451 Librarians 5 1 1 1 1 1

2452 Archivists and curators 4 1 1 1 1 0

3121 Architectural technologists and 4 1 1 1 1 YesTown Planning technicians

3411 Artists 5 1 1 1 1 1 Yes

3412 Authors, Writers 5 1 1 1 1 1 Yes

3413 Actors, Entertainers 5 1 1 1 1 1 Yes

3414 Dancers and Choreographers 5 1 1 1 1 1 Yes

3415 Musicians 5 1 1 1 1 1 Yes

3416 Arts officers, producers and directors 5 1 1 1 1 1 Yes

3417 Photographers, audio-visual and 5 1 1 1 1 1 Yesbroadcasting equipment operators

3421 Graphic Designers 5 1 1 1 1 1 Yes

3422 Product, Clothing and related designers 5 1 1 1 1 1 Yes

3431 Journalists, Newspaper and 5 1 1 1 1 1 YesPeriodical editors

3432 Broadcasting associate professionals 5 1 1 1 1 1 Yes

3433 Public Relations officers 4 1 1 1 1 Yes

3434 Photographers and Audio-Visual 5 1 1 1 1 1 Yesequipment operators

GridScore

Processnovelty

ResistanttoMechanisation

Non-repeatingoutput

CreativeFunctioninprocess

In

terpretationnottransformation

DCMSCreative

TABLE 5.1:OCCUPATIONS WITH A SCORE OF 4 OR 5, WHICH ARE INCLUDED INTHE GRID CLASSIFICATION OF CREATIVE OCCUPATIONS


28/79


3543 Marketing associate professionals 4 1 1 1 1 Yes

5491 Glass and ceramics makers, 5 1 1 1 1 1 Yesdecorators and finishers

5495 Goldsmiths, Silversmiths, 5 1 1 1 1 1 YesPrecious Stone workers

TABLE 5.2: OCCUPATIONS WITH A SCORE OF 1-3, WHICH ARE EXCLUDED FROMTHE GRID DEFINITION OF CREATIVE OCCUPATIONS

SOCcode Occupation

5244 TV, Video and Audio engineers 3 1 1 1 Yes

5422 Printers 3 1 1 1 Yes

5424 Screen printers 3 1 1 1 Yes

5493 Pattern makers (moulds) 3 1 1 1 Yes

5411 Weavers and Knitters 2 1 1 Yes

5496 Floral arrangers, Florists 2 1 1 Yes

8112 Glass and Ceramics process operatives 2 1 1 Yes

5421 Originators, Compositors and 1 1 Yes

Print preparers

5423 Bookbinders and Print finishers 1 1 Yes

5499 Hand Craft occupations not 1 1 Yeselsewhere classified

9121 Labourers in Building and 1 1 YesWoodworking Trades

GridScore

Processnovelty

ResistanttoMechanisation

Non-repeatingoutput

CreativeFunctioninprocess

Interpretationnottransformation

DCMSCreative


29/79


TABLE 5.3:EMPLOYMENT IN OCCUPATIONS GRID-SCORED AS CREATIVE WHICHARE NOT INCLUDED IN THE DCMS DEFINITION

TABLE 5.4:EMPLOYMENT IN OCCUPATIONS INCLUDED IN THE DCMSDEFINITION WHICH ARE NOT GRID-SCORED AS CREATIVE

Code Description Employment

1132 Marketing and sales directors 549,400

2132 Software professionals 327,500

2451 Librarians 28,200

2452 Archivists and curators 11,700

TOTAL 916,800

Code Description Employment

1136 Information and Communication Technology managers 309,900

2126 Design and Development engineers 63,300

5244 TV, Video and Audio engineers 11,400

5411 Weavers and Knitters 2,900

5421 Originators, Compositors and Print preparers 3,500

5422 Printers 33,000

5423 Bookbinders and Print finishers 19,000

5424 Screen printers 1,800

5492 Furniture makers, other craft woodworkers 49,000

5493 Pattern makers (moulds) 1,600

5494 Musical Instrument makers and tuners 2,000

5496 Floral arrangers, Florists 11,900

5499 Hand Craft occupations not elsewhere classified 15,000

8112 Glass and Ceramics process operatives 7,600

9121 Labourers in Building and Woodworking Trades 165,400

Total 698,000

It can been seen from Tables 5.3 and 5.4 that the grid-scoring increases the estimate of

creatively-occupied jobs by 218,800 (916,800-698,000), after rounding, when compared

with the last-published DCMS occupation codes (DCMS 2010:23).


30/79


6. STEP THREE: DERIVING RIGOROUS

INTENSITY MEASURES

We can now apply these grid-generated occupations to generate a new list of creative

intensities for the different industries. We will refer to this as grid intensity where the need

for clarity arises. Using the new occupational definitions (which therefore now deviate from

DCMS to the extent shown in Tables 5.3 and 5.4) we can partition all SIC codes into two

groups on the basis of DCMSs choice of industries.

CHART 6.1: DISTRIBUTION OF CREATIVELY-OCCUPIED JOBS BY GRID INTENSITY,PARTITIONED INTO DCMSCREATIVE AND NONDCMSCREATIVE

This gives a new table of intensities, reproduced in detail in Annex B and illustrated in

Chart 6.1. This is the first step in identifying a baseline set of creative industries. We now

analyse grid-intensities within the DCMScreative industries; we identify the anomolies, and

we then correct them, arriving at a new set of industries defined by their creative intensity.

150,000

200,000

250,000

400,000

300,000

450,000

350,000

100,000

50,000

500,000

550,000

0

00-05 05-15 15-25 25-35 35-45 45-55



occupiedjobs

55-65 65-75 75-85 85-95




31/79


For clarity we refer to the original set of DCMScreative industries as the seed and the

final set as the baseline.8 In Section 8, we test the sensitivity of this method to a different

choice of seed and to a different choice of grid occupations.

Chart 6.1 shows that creativelyoccupied jobs within this seed list of industries falls on a

nearly unimodal distribution, with a mean of 51 per cent, a standard deviation of 19 per cent

and a median of 58 per cent. The distribution of creativelyoccupied jobs within the non

DCMScreative industries has a mean of 5 per cent, a median of 10 per cent and a standard

deviation of 22 per cent. These distribution parameters are so far apart that it is highly

improbable that the two sets of data come from the same population. The interpretation

is clear: there are two distinct groups of industries involved and, equally clearly, some

industries that belong to one group have been misallocated to another.

What should the threshold intensity for a creative industry be? The data do not easily

support imposing a simplifying assumption, for example that they are drawn from normallydistributed populations. But it is obvious, pragmatically, that two distributions are involved

and there is an allocation problem to be settled. We therefore adopt a heuristic device: a

decision-making procedure rooted in the basics of probability theory which can be used

with a range of prior partitions of the data into groups that are assumed to be creative or

not creative, and which eliminates or significantly reduces the inconsistencies in the prior

seed partition by eliminating improbable classifications. This leads to a new partition of

the data which better discriminates between creative and non-creative industries.

An intuitive decision principle is to seek an equal likelihood of a type I error (wrongly

defining a creative industry as non-creative) and a type II error (wrongly defining a non

creative industry as creative). On this basis we set the threshold so that it lies an equal

number of standard deviations from the mean of the distributions. This threshold, it turns

out, falls at 30 per cent on the basis of the DCMSs 2011 creative industry classification,9

which is roughly one standard deviation away from the mean of each group of codes. Any

SIC from the noncreative group which is over this limit is provisionally reclassified as

creative, and any SIC from the creative group which is below it is reclassified as not. This

initial reclassification is finally refined in the next Section by removing a small number of

codes for which the statistical evidence is insufficiently reliable.

Of course this is not the only possible heuristic: we could for example place a greater

weight on the existing DCMS classification by having a lower threshold for exclusion and

a higher threshold for inclusion. In this way we might choose to bias in favour of inclusion,

or to bias in favour of exclusion. Another possibility is to set two different thresholds; onewhich is used to move codes initially assumed to be creative out of that classification if

their intensity is too low, whilst the other is used to move codes initially assumed not to

be creative out of that classification if their intensity is too high. Our choice, which uses

a single equi-probable threshold, has the Bayesian advantage that it imposes the least

assumptions on the data, and this is why we have adopted it. A fruitful topic for research

would be to identify robust and consistent heuristics for partitioning data, like the data for

the creative industries, which have distributions of the type seen in Chart 6.1.

Further options are to analyse the use of ICT labour in greater detail, and finally to take

into consideration other aspects of why an industry may be deemed creative (such as the

nature of its outputs or its production processes), as suggested by our economic model.

These are topics for further research. Our studies so far show, however, that intensity is an

exceptionally good indicator of all other aspects of an industrys creativity, and certainly,

strong doubt must be cast on any choice of creative industries for which intensity is low, or

the exclusion of any industry for which intensity is high.


32/79


CHART 6.2:DISTRIBUTION OF CREATIVELY-OCCUPIED JOBS BY GRIDINTENSITY, PARTITIONED INTO CREATIVE AND NONCREATIVE

The impact of the resulting list of codes is shown in Chart 6.2. It gives a new assignment of

codes to creative industries, with a mean of 57 per cent and standard deviation of 15 per

cent, and non-creative industries with a mean of 4 per cent and a standard deviation of 9 per

cent.

300,000

200,000

100,000

400,000

500,000

600,000

0


Number of

creatively-

occupied jobs

0-5 5-15 15-25 25-35 35-45 45-55 55-65 65-75 75-85 85-95

NonGridCreative Industries GridCreative Industries


33/79


7. STEP FOUR: A STATISTICALLY

RESILIENT BASELINE

In this Section, we refine the baseline to remove statistically volatile or unreliable codes,

and derive the baseline estimate for creative employment arising from this analysis.

The final resulting selection of industries is shown in Tables 7.1 and 7.2 and the baseline

employment estimate in Table 7.3.

Statistical reliability must be taken into account. If the baseline depends on data which

cannot be relied on, then the results may fluctuate erratically, and for reasons not

connected to the underlying nature of the industry. To avoid this, in this Section wedistinguish between industries whose intensity clearly does place them inside, or outside,

the creative industries, and those for which the data is less statistically reliable.

As a rule of thumb, the ONSs Labour Force Survey team advises that individual

employment totals lower than 800 should not be relied on statistically. More technically,

confidence intervals can usually be obtained for estimates based on APS data. Some

thought is needed when applying this information. We do not imply, if we exclude a code

from the baseline on the grounds of statistical reliability, that we are certain it does not

belong there. We are simply saying that the data does not tell us enough to put it there

with confidence.

This procedure may be thought of as conservative in that it is cautious about reclassifying

industries as creative which have not hitherto been thought of as creative, but which seem

so from the intensity analysis.

The headline estimates of the size of the creative industries and creative economy are not

highly sensitive to this procedure. The affected codes, by their very nature, account for

only a small proportion of total creative industry employment. Conclusions drawn from

trends, the weight of the industries in the economy, or the economy-wide composition of

the creative workforce, can therefore be relied on, in the sense that they will be unaffected

by these exclusions. Nevertheless, care is needed for any further conclusion which other

researchers might draw, if that involves small samples containing these undecided codes. In

this paper, we avoid drawing any such conclusions.

A similar consideration leads us to compare the selection of codes for 2010 with a selection

for 2009, to see how much variation occurs between the two years. This also constitutes

an initial test of robustness: we cannot be confident in conclusions that are highly sensitive

to the year from which the data is drawn, unless we can devise a smoothing or aggregation

procedure such as averaging over a number of years. This is the procedure we have

adopted for 20042008, where the baseline classification is constructed from a fouryear

average of creative intensities since data are available on a comparable basis for all of

these years.

For the years 20092010 the only years for which SIC2007 data are available we also

tried to find out which SIC codes were volatilewhen calculated on the basis of a single

years data, that is, those codes whose intensity changed a lot between the two years.


34/79


A useful objective, in further improvements to the DCMS estimates, would be to attach

confidence intervals to the numbers in the published estimates.

For practical purposes we treat as volatile any SIC code that moves from creative to non-creative or vice versa, and which changes by more than one-fifth relative to its lowest

value, between years. Table 7.1, the baseline, excludes reclassified codes which are either

based on small samples, or are volatile, or both. Table 7.2, as noted, presents codes which

have been excluded from the baseline on the above bases, but which, on grounds of

creative intensity alone, might reasonably be included within it.

Table 7.3, finally, presents our baseline estimates of creative employment for 2010. These

combine the occupation codes arising from the Creative Grid with the industrial codes

selected in this Section.

TABLE 7.1:THE BASELINE: CODES DEFINITELY RECOGNISED AS GRIDINTENSIVELY CREATIVE AFTER REMOVING STATISTICALLYUNRELIABLE RECLASSIFICATIONS

Code Description

3212 Manufacture of 6 3 6 4 46% 59% Y Yjewellery andrelated articles

5811 Book publishing 40 17 37 17 42% 46% Y Y Y

5813 Publishing of 512 22 51 19 44% 38% Y Y Ynewspapers

5814 Publishing of 45 29 45 28 63% 62% Y Y Yjournals andperiodicals

5829Other software 22 13 22 13 58% 60% Y Y Ypublishing

5911 Motion picture, 46 31 56 38 69% 68% Y Y Yvideo andtelevisionprogrammeproduction

activities

Comment2

009Total(000)

2009Creative(000)

2010Tota

l(000)

2010Crea

tive(000)

2009Inte

nsity

2010Intensity

2009Grid

-creative

2010Grid

-creative

DCMScre

ative

Smallsam

ple

Volatile

NOTE: Totals in this table are given in thousands to ensure compliance with LFS disclosure requirements.


35/79


5912 Motion picture, 7 5 4 3 71% 83% Y Y Yvideo andtelevision

programmepost-productionactivities

6010 Radio 14 10 16 13 69% 79% Y Y Ybroadcasting

6020 Television 43 22 39 22 51% 58% Y Y Yprogrammingand broadcastingactivities

6201 Computer 195 110 207 120 56% 58% Y Yprogramming

activities

6202Computer 233 125 202 112 54% 55% Y Yconsultancyactivities

6209 Other infor- 34 14 35 12 40% 36% Y Ymation tech-nology andcomputerservice activities

7021 Public relations 30 21 27 18 72% 67% Y Y

and communi-cation activities

7111 Architectural 93 60 96 63 64% 65% Y Y Yactivities

7311 Advertising 85 49 87 45 58% 52% Y Y Yagencies

7312 Media 31 18 24 14 56% 57% Y Y Yrepresentation

7320 Market research 39 12 42 15 30% 35% Yand public

opinion polling7410 Specialised 101 66 105 61 65% 58% Y Y Y

design activities

7420 Photographic 44 31 41 30 71% 75% Y Y Yactivities

7430 Translation and 17 11 14 10 66% 74% Y Yinterpretationactivities

9001 Performing arts 39 29 45 36 74% 80% Y Y Y

9002 Support activities 10 6 11 6 58% 54% Y Y Y

to performing arts

9003 Artistic creation 71 67 71 63 95% 89% Y Y Y


36/79


2341 Manufacture of 6 2 7 4 42% 57% Y Y Yceramichousehold and

ornamentalarticles

5821 Publishing of 3 1 2 1 53% 38% Y Y Y Y Ycomputergames

5819 Other 32 13 37 12 41% 32% Y Y Y Ypublishingactivities

5913 Motion picture, 4 1 9 3 28% 35% Y Y Yvideo andtelevisionprogrammedistributionactivities

5920 Sound recording 15 10 10 5 68% 51% Y Y Y Yand musicpublishingactivities

1820 Reproduction 8 3 6 4 40% 64% Y Y Y Yof recordedmedia

Volatile,

not part ofthe DCMSdefinition; but

high intensityin both years.

Included.

Volatile and asmall sample,but part ofthe DCMSdefinition.

Included.

Volatile, butlarge sampleand part ofthe DCMS

definition.Included.

Volatile, not a

large sample,

but part of the

DCMS definition,

close to the

2009 threshold

and above the

2010 threshold.

Included.

Volatile, but

large sample

and part of the

DCMS definition.

Included.

Volatile, not

large sample,

but part of the

DCMS definition.

Included.


37/79


Code Description Comment

2640Manufacture of 7 2 5 2 37% 31% Y Yconsumer

electronics

6120 Wireless 83 18 83 26 22% 31% Y Ytelecom-municationsactivities

2342 Manufacture 4 1 2 1 28% 47% Y Yceramic sanitaryfixtures

2009Total

2009Creative

2010Total

2010Creative

2009Intensity

2010Intensity

2009Grid-creative

2010Grid-creative

DCMScreative

Smallsample

Volatile

Borderline

case, not large

sample and near

the threshold.Not part of the

DCMS definition.

Excluded, but

a plausible

candidate for

inclusion.

Large sample,

but volatile, near

the threshold,

not part of the

DCMS definition.

Excluded, but

a plausible

candidate for

inclusion.

Volatile, and

not part of the

DCMS definition.

Excluded.

TABLE 7.2:CODES THAT ARE SUGGESTED BY THE CREATIVE INTENSITYANALYSIS, BUT EXCLUDED ON GROUNDS OF INSUFFICIENTSTATISTICAL RELIABILITY

BASELINE EMPLOYMENT ESTIMATES

Table 7.3 presents our baseline estimates of creative employment, derived as noted

by combining the grid-selected occupations with the industries identified as creative

according to their intensities, as modified by the restrictions of statistical reliability

imposed in this Section.

In line with the Creative Trident methodology introduced by Higgs et al. (2005, 2008), we

use the term specialist to refer to workers who are creatively occupied and work within

the creative industries; support workers refers to workers who are not creatively occupied,

but work within the creative industries; and embedded workers are creatively occupied

outside the creative industries.

TABLE 7.3:BASELINE EMPLOYMENT ESTIMATES

Specialist Support Embedded Total

794,000 563,300 1,138,400 2,495,700


38/79


8. TESTING THE GRID:

REVERSE INTENSITY AND THESPECIALISATION OF EMPLOYMENTIN THE CREATIVE INDUSTRIES

In the next Section we will test the sensitivity of the employment estimates to our

assumptions. Before doing so, we conduct a brief further reality check on our economic

model using what we term reverse intensity (Freeman (2007) terms this occupational

intensity). We define reverse intensity as the proportion of the total employment of agiven occupation that is found within a given industry. In contrast, normal intensity is

the proportion of total employment of a given industry that is accounted for by a given

occupation.

Occupations with high reverse intensities tend to be specialised so that architects, for

example, have a high reverse intensity within the architecture sector where most of them

work, and correspondingly low reverse intensities elsewhere. If our model is correct, then

not only should the creative industries be intensive employers of creative occupations, but

in addition these creative occupations should be found in greater concentrations within

these industries. To take a concrete example, it is not only the case that the architecture

sector uses many architects, but also that many architects work in the architecture sector.

These two statements may sound as if they are two ways of saying the same thing, but

they are not. It could be, for example (though this is not the case), that only 5 per cent

of architects work within the architecture industry, whilst 85 per cent of the workforce of

those industries is made up of architects.

We test the claim that the creative occupations are a specialist resource, which the creative

industries make especial use of, by calculating the reverse intensities of the occupations

that we treat as creative in our baseline. The results are shown in Table 8.1

TABLE 8.1: REVERSE INTENSITY: PROPORTION OF EACH GRID-DEFINED

CREATIVE OCCUPATION WHICH WORKS WITHIN THE BASELINEINDUSTRIES

Code Description Rev. Intensity

3432 Broadcasting associate professionals 89%

3411 Artists 82%

3431 Journalists, Newspaper and Periodical editors 78%

2431 Architects 75%

3412 Authors, Writers 74%

3434 Photographers and Audio-Visual equipment operators 71%

3421 Graphic Designers 70%

3421 Arts officers, producers and directors 64%


39/79


3415 Musicians 59%

3121 Architectural technologists and Town Planning technicians 56%

3422 Product, Clothing and related designers 56%

2131 IT Strat

Documents

A Dynamic Mapping of the Creative Industries